gpumd

Calculations

thermo.gpumd.calc.get_gkma_kappa(data, nbins, nsamples, dt, sample_interval, T=300, vol=1, max_tau=None, directions='xyz', outputfile='heatmode.npy', save=False, directory=None, return_data=True)

Calculate the Green-Kubo thermal conductivity from modal heat current data from ‘load_heatmode’

Args:

data (dict):

Dictionary with heat currents loaded by ‘load_heatmode’

nbins (int):

Number of bins used during the GPUMD simulation

nsamples (int):

Number of times heat flux was sampled with GKMA during GPUMD simulation

dt (float):

Time step during data collection in fs

sample_interval (int):

Number of time steps per sample of modal heat flux

T (float):

Temperature of system during data collection

vol (float):

Volume of system in angstroms^3

max_tau (float):

Correlation time to calculate up to. Units of ns

directions (str):

Directions to gather data from. Any order of ‘xyz’ is accepted. Excluding directions also allowed (i.e. ‘xz’ is accepted)

outputfile (str):

File name to save read data to. Output file is a binary dictionary. Loading from a binary file is much faster than re-reading data files and saving is recommended

save (bool):

Toggle saving data to binary dictionary. Loading from save file is much faster and recommended

directory (str):

Name of directory storing the input file to read

return_data (bool):

Toggle returning the loaded modal heat flux data. If this is False, the user should ensure that save is True

Returns:

dict: Input data dict but with correlation, thermal conductivity, and lag time data included

Output dictionary (new entries)

key	tau	kmxi	kmxo	kmyi	kmyo	kmz	jmxijx	jmxojx	jmyijy	jmyojy	jmzjz
units	ns	Wm-1 K-1 x-1	Wm-1 K-1 x-1	Wm-1 K-1 x-1	Wm-1 K-1 x-1	Wm-1 K-1 x-1	eV3 amu-1 x-1	eV3 amu-1 x-1	eV3 amu-1 x-1	eV3 amu-1 x-1	eV3 amu-1 x-1

Here x is the size of the bins in THz. For example, if there are 4 bins per THz, x = 0.25 THz.

thermo.gpumd.calc.hnemd_spectral_kappa(shc, Fe, T, V)

Spectral thermal conductivity calculation from an SHC run

Args:

shc (dict):

The data from a single SHC run as output by thermo.gpumd.data.load_shc

Fe (float):

HNEMD force in (1/A)

T (float):

HNEMD run temperature (K)

V (float):

Volume (A^3) during HNEMD run

Returns:

dict: Same as shc argument, but with spectral thermal conductivity included

Output dictionary (new entries)

key	kwi	kwo
units	Wm-1 K-1 THz-1	Wm-1 K-1 THz-1

thermo.gpumd.calc.running_ave(kappa, time)

Gets running average. Reads and returns the structure input file from GPUMD.

Args:

kappa (ndarray): Raw thermal conductivity time (ndarray): Time vector that kappa was sampled at

Returns:

ndarray: Running average of kappa input

Data Loaders

thermo.gpumd.data.get_frequency_info(bin_f_size, eigfile='eigenvector.out', directory=None)

Gathers eigen-frequency information from the eigenvector file and sorts it appropriately based on the selected frequency bins (identical to internal GPUMD representation).

Args:

bin_f_size (float):

The frequency-based bin size (in THz)

eigfile (str):

The filename of the eigenvector output/input file created by GPUMD phonon package

directory (str):

Directory eigfile is stored

Returns:

dict: Dictionary with the system eigen-freqeuency information along with binning information

Output dictionary

key	fq	fmax	fmin	shift	nbins	bin_count	bin_f_size
units	THz	THz	THz	N/A	N/A	N/A	THz

thermo.gpumd.data.load_compute(quantities=None, directory=None, filename='compute.out')

Loads data from compute.out GPUMD output file.

Currently supports loading a single run.

Args:

quantities (str or list(str)):

Quantities to extract from compute.out Accepted quantities are:

[‘T’, ‘U’, ‘F’, ‘W’, ‘jp’, ‘jk’].

Other quantity will be ignored.

T=temperature, U=potential, F=force, W=virial, jp=heat current (potential), jk=heat current (kinetic)

directory (str):

Directory to load compute file from

filename (str):

file to load compute from

Returns:

Dictionary containing the data from compute.out

Output dictionary

key	T	U	F	W	jp	jk	Ein	Eout
units	K	eV	eVA-1	eV	eV3/2 amu-1/2	eV3/2 amu-1/2	eV	eV

thermo.gpumd.data.load_dos(num_dos_points, directory=None, filename='dos.out')

Loads data from dos.out GPUMD output file.

Args:

num_dos_points (int or list(int)):

Number of frequency points the DOS is computed for.

directory (str):

Directory to load ‘dos.out’ file from (dir. of simulation)

filename (str):

File to load DOS from.

Returns:

dict(dict)): Dictonary with DOS data. The outermost dictionary stores each individual run.

Output dictionary

key	nu	DOSx	DOSy	DOSz
units	THz	THz-1	THz-1	THz-1

thermo.gpumd.data.load_force(n, directory=None, filename='force.out')

Loads data from force.out GPUMD output file.

Currently supports loading a single run.

Args:

n (int):

Number of atoms force is output for

directory (str):

Directory to load force file from

filename (str):

Name of force data file

Returns:

Numpy array of shape (n,3,-1) containing all forces (ev/A) from filename

thermo.gpumd.data.load_hac(Nc, output_interval, directory=None, filename='hac.out')

Loads data from hac.out GPUMD output file.

Args:

Nc (int or list(int)):

Number of correlation steps

output_interval (int or list(int)):

Output interval for HAC and RTC data

directory (str):

Directory containing hac data file

filename (str):

The hac data file

Returns:

dict: A dictionary containing the data from hac runs

Output dictionary

key	t	kxi	kxo	kyi	kyo	kz	jxijx	jxojx	jyijy	jyojy	jzjz
units	ps	Wm-1 K-1	Wm-1 K-1	Wm-1 K-1	Wm-1 K-1	Wm-1 K-1	eV3 amu-1	eV3 amu-1	eV3 amu-1	eV3 amu-1	eV3 amu-1

thermo.gpumd.data.load_heatmode(nbins, nsamples, directory=None, inputfile='heatmode.out', directions='xyz', outputfile='heatmode.npy', ndiv=None, save=False, multiprocessing=False, ncore=None, block_size=65536, return_data=True)

Loads data from heatmode.out GPUMD file. Option to save as binary file for fast re-load later. WARNING: If using multiprocessing, memory usage may be significantly larger than file size

Args:

nbins (int):

Number of bins used during the GPUMD simulation

nsamples (int):

Number of times heat flux was sampled with GKMA during GPUMD simulation

directory (str):

Name of directory storing the input file to read

inputfile (str):

Modal heat flux file output by GPUMD

directions (str):

Directions to gather data from. Any order of ‘xyz’ is accepted. Excluding directions also allowed (i.e. ‘xz’ is accepted)

outputfile (str):

File name to save read data to. Output file is a binary dictionary. Loading from a binary file is much faster than re-reading data files and saving is recommended

ndiv (int):

Integer used to shrink number of bins output. If originally have 10 bins, but want 5, ndiv=2. nbins/ndiv need not be an integer

save (bool):

Toggle saving data to binary dictionary. Loading from save file is much faster and recommended

multiprocessing (bool):

Toggle using multi-core processing for conversion of text file

ncore (bool):

Number of cores to use for multiprocessing. Ignored if multiprocessing is False

block_size (int):

Size of block (in bytes) to be read per read operation. File reading performance depend on this parameter and file size

return_data (bool):

Toggle returning the loaded modal heat flux data. If this is False, the user should ensure that save is True

Returns:

dict: Dictionary with all modal heat fluxes requested

Output dictionary

key	nbins	nsamples	jmxi	jmxo	jmyi	jmyo	jmz
units	N/A	N/A	eV3/2 amu-1/2 x-1	eV3/2 amu-1/2 x-1	eV3/2 amu-1/2 x-1	eV3/2 amu-1/2 x-1	eV3/2 amu-1/2 x-1

Here x is the size of the bins in THz. For example, if there are 4 bins per THz, x = 0.25 THz.

thermo.gpumd.data.load_kappa(directory=None, filename='kappa.out')

Loads data from kappa.out GPUMD output file which contains HNEMD kappa.

Args:

directory (str):

Directory containing kappa data file

filename (str):

The kappa data file

Returns:

dict: A dictionary with keys corresponding to the columns in ‘kappa.out’

Output dictionary

key	kxi	kxo	kyi	kyo	kz
units	Wm-1 K-1	Wm-1 K-1	Wm-1 K-1	Wm-1 K-1	Wm-1 K-1

thermo.gpumd.data.load_kappamode(nbins, nsamples, directory=None, inputfile='kappamode.out', directions='xyz', outputfile='kappamode.npy', ndiv=None, save=False, multiprocessing=False, ncore=None, block_size=65536, return_data=True)

Loads data from kappamode.out GPUMD file. Option to save as binary file for fast re-load later. WARNING: If using multiprocessing, memory useage may be significantly larger than file size

Args:

nbins (int):

Number of bins used during the GPUMD simulation

nsamples (int):

Number of times thermal conductivity was sampled with HNEMA during GPUMD simulation

directory (str):

Name of directory storing the input file to read

inputfile (str):

Modal thermal conductivity file output by GPUMD

directions (str):

Directions to gather data from. Any order of ‘xyz’ is accepted. Excluding directions also allowed (i.e. ‘xz’ is accepted)

outputfile (str):

File name to save read data to. Output file is a binary dictionary. Loading from a binary file is much faster than re-reading data files and saving is recommended

ndiv (int):

Integer used to shrink number of bins output. If originally have 10 bins, but want 5, ndiv=2. nbins/ndiv need not be an integer

save (bool):

Toggle saving data to binary dictionary. Loading from save file is much faster and recommended

multiprocessing (bool):

Toggle using multi-core processing for conversion of text file

ncore (bool):

Number of cores to use for multiprocessing. Ignored if multiprocessing is False

block_size (int):

Size of block (in bytes) to be read per read operation. File reading performance depend on this parameter and file size

return_data (bool):

Toggle returning the loaded modal thermal conductivity data. If this is False, the user should ensure that save is True

Returns:

dict: Dictionary with all modal thermal conductivities requested

Output dictionary

key	nbins	nsamples	kmxi	kmxo	kmyi	kmyo	kmz
units	N/A	N/A	Wm-1 K-1 x-1	Wm-1 K-1 x-1	Wm-1 K-1 x-1	Wm-1 K-1 x-1	Wm-1 K-1 x-1

Here x is the size of the bins in THz. For example, if there are 4 bins per THz, x = 0.25 THz.

thermo.gpumd.data.load_omega2(directory=None, filename='omega2.out')

Loads data from omega2.out GPUMD output file.

Args:

directory (str):

Directory to load force file from

filename (str):

Name of force data file

Returns:

Numpy array of shape (N_kpoints,3*N_basis) in units of THz. N_kpoints is number of k points in kpoint.in and N_basis is the number of basis atoms defined in basis.in

thermo.gpumd.data.load_saved_heatmode(filename='heatmode.npy', directory=None)

Loads data saved by the ‘load_heatmode’ or ‘get_gkma_kappa’ function and returns the original dictionary.

Args:

filename (str):

Name of the file to load

directory (str):

Directory the data file is located in

Returns:

dict: Dictionary with all modal heat flux previously requested

thermo.gpumd.data.load_saved_kappamode(filename='kappamode.npy', directory=None)

Loads data saved by the ‘load_kappamode’ function and returns the original dictionary.

Args:

filename (str):

Name of the file to load

directory (str):

Directory the data file is located in

Returns:

dict: Dictionary with all modal thermal conductivities previously requested

thermo.gpumd.data.load_sdc(Nc, directory=None, filename='sdc.out')

Loads data from sdc.out GPUMD output file.

Args:

Nc (int or list(int)):

Number of time correlation points the VAC/SDC is computed for

directory (str):

Directory to load ‘sdc.out’ file from (dir. of simulation)

filename (str):

File to load SDC from

Returns:

dict(dict):

Dictonary with SDC/VAC data. The outermost dictionary stores each individual run

Output dictionary

key	t	VACx	VACy	VACz	SDCx	SDCy	SDCz
units	ps	A2 ps-2	A2 ps-2	A2 ps-2	A2 ps-1	A2 ps-1	A2 ps-1

thermo.gpumd.data.load_shc(Nc, num_omega, directory=None, filename='shc.out')

Loads the data from shc.out GPUMD output file.

Args:

Nc (int or list(int)):

Maximum number of correlation steps. If multiple shc runs, can provide a list of Nc.

num_omega (int or list(int)):

Number of frequency points. If multiple shc runs, can provide a list of num_omega.

directory (str):

Directory to load ‘shc.out’ file from (dir. of simulation)

filename (str):

File to load SHC from.

Returns:

dict: Dictionary of in- and out-of-plane shc results (average)

Output dictionary

key	t	Ki	Ko	nu	jwi	jwo
units	ps	A eV ps-1	A eV ps-1	THz	A eV ps-1 THz-1	A eV ps-1 THz-1

thermo.gpumd.data.load_thermo(directory=None, filename='thermo.out')

Loads data from thermo.out GPUMD output file.

Args:

directory (str):

Directory to load thermal data file from

filename (str):

Name of thermal data file

Returns:

‘output’ dictionary containing the data from thermo.out

Output dictionary

key	T	K	U	Px	Py	Pz	Lx	Ly	Lz	ax	ay	az	bx	by	bz	cx	cy	cz
units	K	eV	eV	GPa	GPa	GPa	A	A	A	A	A	A	A	A	A	A	A	A

thermo.gpumd.data.load_vac(Nc, directory=None, filename='mvac.out')

Loads data from mvac.out GPUMD output file.

Args:

Nc (int or list(int)):

Number of time correlation points the VAC is computed for

directory (str):

Directory to load ‘mvac.out’ file from

filename (str):

File to load VAC from

Returns:

dict(dict):

Dictonary with VAC data. The outermost dictionary stores each individual run

Output dictionary

key	t	VACx	VACy	VACz
units	ps	A2 ps-2	A2 ps-2	A2 ps-2

thermo.gpumd.data.load_velocity(n, directory=None, filename='velocity.out')

Loads data from velocity.out GPUMD output file.

Currently supports loading a single run.

Args:

n (int):

Number of atoms velocity is output for

directory (str):

Directory to load velocity file from

filename (str):

Name of velocity data file

Returns:

Numpy array of shape (n,3,-1) containing all forces (A/ps) from filename

thermo.gpumd.data.reduce_frequency_info(freq, ndiv=1)

Recalculates frequency binning information based on how many times larger bins are wanted.

Args:

freq (dict): Dictionary with frequency binning information from the get_frequency_info function output

ndiv (int):

Integer used to shrink number of bins output. If originally have 10 bins, but want 5, ndiv=2. nbins/ndiv need not be an integer

Returns:

dict: Dictionary with the system eigen freqeuency information along with binning information

thermo.gpumd.data.tail(f, nlines, BLOCK_SIZE=32768)

Reads the last nlines of a file.

Args:

f (filehandle):

File handle of file to be read

nlines (int):

Number of lines to be read from end of file

BLOCK_SIZE (int):

Size of block (in bytes) to be read per read operation. Performance depend on this parameter and file size.

Returns:

list: List of ordered final nlines of file

Additional Information: Since GPUMD output files are mostly append-only, this becomes useful when a simulation prematurely ends (i.e. cluster preempts run, but simulation restarts elsewhere). In this case, it is not necessary to clean the directory before re-running. File outputs will be too long (so there still is a storage concern), but the proper data can be extracted from the end of file. This may also be useful if you want to only grab data from the final m number of runs of the simulation

Input/Output

thermo.gpumd.io.ase_atoms_to_gpumd(atoms, M, cutoff, gpumd_file='xyz.in', sort_key=None, order=None, group_index=None)

Converts ASE atoms to GPUMD compatible position file.

Args:

atoms (ase.Atoms):

Atoms to write to gpumd file

M (int):

Maximum number of neighbors for one atom

cutoff (float):

Initial cutoff distance for building the neighbor list

gpumd_file (str):

File to save the structure data to

sort_key (str):

How to sort atoms (‘group’, ‘type’).

order (list(type)):

List to sort by. Provide str for ‘type’, and int for ‘group’

group_index (int):

Selects the group to sort in the output.

thermo.gpumd.io.convert_gpumd_atoms(in_file='xyz.in', out_filename='in.xyz', format='xyz', atom_types=None)

Converts the GPUMD input structure file to any compatible ASE output structure file. Warning: Info dictionary may not be preserved.

Args:

in_file (str):

GPUMD position file to get structure from

out_filename (str):

Name of output file after conversion

format (str):

ASE supported output format

atom_types (list(str)):

List of atom types (elements).

thermo.gpumd.io.create_basis(atoms)

Creates the basis.in file. Atoms passed to this must already have the basis of every atom defined.

Related: preproc.add_basis, preproc.repeat

Args:

atoms (ase.Atoms):

Atoms of unit cell used to generate basis.in

thermo.gpumd.io.create_kpoints(atoms, path='G', npoints=1, special_points=None)

Creates the file “kpoints.in”, which specifies the kpoints needed for src/phonon

Args:

atoms (ase.Atoms):

Unit cell to use for phonon calculation

path (str):

String of special point names defining the path, e.g. ‘GXL’

npoints (int):

Number of points in total. Note that at least one point is added for each special point in the path

special_points (dict):

Dictionary mapping special points to scaled kpoint coordinates. For example {'G': [0, 0, 0], 'X': [1, 0, 0]}

Returns:

tuple: First element is the kpoints converted to x-coordinates, second the x-coordinates of the high symmetry points, and third the labels of those points.

thermo.gpumd.io.import_trajectory(filename='movie.xyz', in_file=None, atom_types=None)

Reads the trajectory from GPUMD run and creates a list of ASE atoms.

Args:

filename (str):

Name of the file that holds the GPUMD trajectory.

in_file (str):

Name of the original structure input file. Not required, but can help load extra information not included in trajectory output.

atom_types (list(str)):

List of atom types (elements).

Returns:

list(ase.Atoms): A list of ASE atoms objects.

thermo.gpumd.io.lammps_atoms_to_gpumd(filename, M, cutoff, style='atomic', gpumd_file='xyz.in')

Converts a lammps data file to GPUMD compatible position file.

Args:

filename (str):

LAMMPS data file name

M (int):

Maximum number of neighbors for one atom

cutoff (float):

Initial cutoff distance for building the neighbor list

style (str):

Atom style used in LAMMPS data file

gpumd_file (str):

File to save the structure data to

thermo.gpumd.io.load_xyz(filename='xyz.in', atom_types=None)

Reads and returns the structure input file from GPUMD.

Args:

filename (str):

Name of structure file

atom_types (list(str)):

List of atom types (elements).

Returns:

tuple: atoms, M, cutoff

atoms (ase.Atoms): ASE atoms object with x,y,z, mass, group, type, cell, and PBCs from input file. group is stored in tag, atom type may not correspond to correct atomic symbol

M (int): Max number of neighbor atoms

cutoff (float): Initial cutoff for neighbor list build

Preprocessing

thermo.gpumd.preproc.add_basis(atoms, index=None, mapping=None)

Assigns a basis index for each atom in atoms. Updates atoms.

Args:

atoms (ase.Atoms):

Atoms to assign basis to.

index (list(int)):

Atom indices of those in the unit cell. Order is important.

mapping (list(int)):

Mapping of all atoms to the relevant basis positions

thermo.gpumd.preproc.add_group_by_position(split, atoms, direction)

Assigns groups to all atoms based on its position. Only works in one direction as it is used for NEMD. Returns a bookkeeping parameter, but atoms will be udated in-place.

Args:

split (list(float)):

List of boundaries. First element should be lower boundary of sim. box in specified direction and the last the upper.

atoms (ase.Atoms):

Atoms to group

direction (str):

Which direction the split will work.

Returns:

int: A list of number of atoms in each group.

thermo.gpumd.preproc.add_group_by_type(atoms, types)

Assigns groups to all atoms based on atom types. Returns a bookkeeping parameter, but atoms will be udated in-place.

Args:

atoms (ase.Atoms):

Atoms to group

types (dict):

Dictionary with types for keys and group as a value. Only one group allowed per atom. Assumed groups are integers starting at 0 and increasing in steps of 1. Ex. range(0,10).

Returns:

int: A list of number of atoms in each group.

thermo.gpumd.preproc.repeat(atoms, rep)

A wrapper of ase.Atoms.repeat that is aware of GPUMD’s basis information.

Args:

atoms (ase.Atoms):

Atoms to assign velocities to.

rep (int | list(3 ints)):

List of three positive integers or a single integer

thermo.gpumd.preproc.set_velocities(atoms, custom=None)

Sets the ‘velocity’ part of the atoms to be used in GPUMD. Custom velocities must be provided. They must also be in the units of eV^(1/2) amu^(-1/2).

Args:

atoms (ase.Atoms):

Atoms to assign velocities to.

custom (list(list)):

list of len(atoms) with each element made from a 3-element list for [vx, vy, vz]